1. Technical Field of the Invention
The present invention relates to a technology that enables provision of 3 dimensional sound with a high feeling of presence or realism to a listener.
2. Description of the Related Art
Examples of a technology for providing 2 or 3-Dimensional (2D or 3D) sound with a high feeling of presence or realism include a so-called multichannel surround system. In the multichannel surround system, multiple speakers, which are arranged around a listener, emit sounds (so as to surround the listener) to provide a 2D or 3D sound with a high sense of presence or realism. International Telecommunication Union (ITU) has made recommendations as to the positions of arrangement of the speakers in such a multichannel surround system (see Non-Patent Reference 1). For example, for a system including 5 speakers, i.e., a center channel speaker C, a left front speaker L, a right front speaker R, a left surround speaker LS, and a right surround speaker RS, it is recommended that the speakers be arranged as shown in FIG. 4. In the following, the left front speaker L and the left surround speaker LS are commonly referred to as a “left channel speaker” or simply “left speaker” when there is no need to discriminate between the two channels and the right front speaker R and the right surround speaker RS are also commonly referred to as a “right channel speaker” or simply “right speaker” when there is no need to discriminate between the two channels.
The left front speaker L which is arranged at a front left side when viewed from the listener and the right front speaker R which is arranged at a front right side as shown in FIG. 4 are used to localize a sound image at the front left side, the center front side, or the front right side from the viewpoint of the listener. The left surround speaker LS and the right surround speaker RS, which will be collectively referred to as “surround speakers” in some cases, are arranged, respectively, at the left lateral side (or left rear side) and the right lateral side (or right rear side) of the listener and are used to reproduce a non-localized sound (for example, a sound such as speech coming out of nowhere) or a localized sound of a sound image of the lateral side or the rear side of the listener. The center channel speaker C arranged at the center front side of the listener is used to reproduce a sound localized at the front side such as a line of dialog of, for example, a drama or movie. A system (so-called 5.1 channel surround system) which includes a subwoofer responsible for mid and bass ranges in addition to the 5 speakers shown in FIG. 4 has also been widely used.
Sounds output from the speakers in the multichannel surround system described above not only include sounds recorded using a general microphone but also frequently include sounds recorded using a so-called dummy head. Accordingly, it is possible to provide a 3D sound with a high sense of presence or realism even though the speakers are arranged in 2 dimensions. Here, the term “dummy head recording” refers to a technology for receiving and recording sounds of microphones arranged respectively at positions of left and right ears of a human head model (i.e., a dummy head). In the following description, an output signal of a microphone at the left ear side of the dummy head is referred to as a “left dummy head signal DL” and an output signal of a microphone at the right ear side thereof is referred to as a “right dummy head signal DR”.
However, a phenomenon which is called “crosstalk” may occur when the left and right speakers are driven by the dummy head signals. Here, the crosstalk is, for example, a phenomenon in which sound emitted from the speaker of the right channel travels around the head of the listener to reach the left ear EL of the listener (or, similarly, a phenomenon in which a sound emitted from the speaker of the left channel travels around the head of the listener to reach the right ear ER of the listener). Thus, a technology in which each dummy head signal is provided to each speaker after preprocessing is performed on the dummy head signal through a filtering process or the like to cancel the crosstalk has been suggested (for example, see Patent Reference 1).